CN107798653A - A kind of method of image procossing and a kind of device - Google Patents
A kind of method of image procossing and a kind of device Download PDFInfo
- Publication number
- CN107798653A CN107798653A CN201710853951.7A CN201710853951A CN107798653A CN 107798653 A CN107798653 A CN 107798653A CN 201710853951 A CN201710853951 A CN 201710853951A CN 107798653 A CN107798653 A CN 107798653A
- Authority
- CN
- China
- Prior art keywords
- image
- feature region
- fisrt feature
- region
- adjustment
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 235000013311 vegetables Nutrition 0.000 claims description 52
- 238000013527 convolutional neural network Methods 0.000 claims description 47
- 230000000694 effects Effects 0.000 claims description 37
- 238000004590 computer program Methods 0.000 claims description 14
- 238000001514 detection method Methods 0.000 claims description 14
- 238000007689 inspection Methods 0.000 claims description 12
- 238000012216 screening Methods 0.000 claims description 12
- 238000001914 filtration Methods 0.000 claims description 11
- 230000008859 change Effects 0.000 claims description 6
- 230000006399 behavior Effects 0.000 claims 2
- 230000004069 differentiation Effects 0.000 abstract description 3
- 230000011218 segmentation Effects 0.000 description 18
- 238000012545 processing Methods 0.000 description 10
- 238000011156 evaluation Methods 0.000 description 9
- 230000006870 function Effects 0.000 description 9
- 210000002569 neuron Anatomy 0.000 description 9
- 235000013305 food Nutrition 0.000 description 8
- 238000012549 training Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 7
- 238000013507 mapping Methods 0.000 description 5
- 235000015277 pork Nutrition 0.000 description 5
- 235000015067 sauces Nutrition 0.000 description 5
- 238000003860 storage Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 235000013399 edible fruits Nutrition 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000003062 neural network model Methods 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 235000009508 confectionery Nutrition 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000007935 neutral effect Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000013519 translation Methods 0.000 description 2
- 241000031708 Saprospiraceae Species 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000000746 body region Anatomy 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 235000012149 noodles Nutrition 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000003507 refrigerant Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000000946 synaptic effect Effects 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 239000011800 void material Substances 0.000 description 1
Classifications
-
- G06T3/04—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
Abstract
The embodiments of the invention provide a kind of a kind of mobile terminal of the method and apparatus of image procossing, methods described is applied to mobile terminal；Described method includes：Client-initiated image acquisition commands are received, gather the first image；Fisrt feature region, and second feature region are determined from described first image；The first adjustment is carried out to the fisrt feature region respectively to operate, and, the second adjustment is carried out to the second feature region and is operated, obtains the second image after having performed the first adjustment operation and the second adjustment operation；Show second image.The embodiment of the present invention can be directed to the characteristics of different zones in same image and carry out differentiation adjustment respectively, rather than unified adjustment is carried out to same image, add the aesthetic feeling of image, improve Consumer's Experience.
Description
Technical field
The present invention relates to information data technical field, more particularly to a kind of image processing method and a kind of image procossing
Device.
Background technology
In recent years, user gradually increases for the demand of U.S. figure, and scene is more and more extensive.
For example, U.S. group and masses' comment have the entrance that user uploads cuisines scene picture, but scheme without corresponding U.S.
Function.User if necessary to be beautified to image, it is necessary to scheme APP by means of other beautiful, then the picture after processing is uploaded,
Consumer's Experience can so be influenceed.
In addition, existing cuisines scene includes a large amount of personages and cuisines while appears in the situation in piece image, and it is traditional
U.S. figure APP generally require user and integrally beautified for image manually, or a certain region that user in image specifies is entered
Row beautification, the effect of so U.S. figure are also not ideal.
The content of the invention
The embodiment of the present invention provides a kind of method of image procossing and corresponding a kind of device of image procossing, a kind of electronics
Equipment, a kind of computer-readable recording medium, to solve user's cumbersome, picture beautification effect during beautifying to picture
The undesirable above mentioned problem of fruit.
In order to solve the above problems, the embodiment of the invention discloses a kind of method of image procossing, methods described is applied to
Mobile terminal, described method include：
Client-initiated image acquisition commands are received, gather the first image；
Fisrt feature region, and second feature region are determined from described first image；
The first adjustment is carried out to the fisrt feature region respectively to operate, and, the is carried out to the second feature region
Two adjustment operations, obtain the second image after having performed the first adjustment operation and the second adjustment operation；
Show second image.
Preferably, the application program specified is installed on the mobile terminal；
It is described reception Client-initiated image capture instruction, gather the first image the step of include：
The image acquisition commands that user is initiated by the application program specified are received, gather the first image；
Described the step of showing second image, includes：
Second image is showed by the application program specified.
Preferably, the fisrt feature region is the region comprising portrait, and the second feature region is to include vegetable
Region.
Preferably, it is described to determine that fisrt feature region, and the step of second feature region are wrapped from described first image
Include：
Using the example parted pattern based on depth convolutional neural networks pre-set, described first image is carried out real
Example segmentation, obtains label corresponding to each pixel, and the type of label；
The region that the type identical pixel of all labels forms is defined as fisrt feature region, and the second spy respectively
Levy region.
Preferably, described the step of the first adjustment operation is carried out to the fisrt feature region, includes：
Recognition of face, and positioning key point are carried out to the fisrt feature region；
Utilize personage in fisrt feature region described in the age detection depth convolutional neural networks model inspection pre-set
Age, and, using in fisrt feature region described in the sex-screening depth convolutional neural networks model inspection pre-set
The sex of personage；
Obtain the special effect parameters that there is corresponding relation with the age, sex；
It is the special effect parameters by the present image parameter change in the fisrt feature region for the key point.
Preferably, described the step of the second adjustment operation is carried out to the second feature region, includes：
The second feature region is identified using the vegetable identification model based on depth convolutional neural networks pre-set
Vegetable title；
Obtain the filter params that there is corresponding relation with the title of the vegetable；
Filtering effects are added to the second feature region using the filter params.
Preferably, the step of obtaining the second image after having performed the first adjustment operation and the second adjustment operation is wrapped
Include：
The personalized sentence that there is corresponding relation with menu name is obtained from preset personalized statement library；
The personalized sentence is added in default key point；
Fisrt feature region, second feature region after being operated based on adjustment, and the personalized sentence, generation second
Image.
In order to solve the above problems, the embodiment of the invention also discloses a kind of device of image procossing, including：
Acquisition module, for receiving Client-initiated image acquisition commands, gather the first image；
Area determination module, for determining fisrt feature region, and second feature region from described first image；
Image adjustment module, operated for carrying out the first adjustment to the fisrt feature region respectively, and, to described the
Two characteristic areas carry out the second adjustment operation, obtain the second figure after having performed the first adjustment operation and the second adjustment operation
Picture；
Display module, for showing second image.
Preferably, the application program specified including is installed on the mobile terminal of described device；
The acquisition module is additionally operable to receive the image acquisition commands that user is initiated by the application program specified, and adopts
Collect the first image；
The display module is additionally operable to show second image by the application program specified.
Preferably, the fisrt feature region is the region comprising portrait, and the second feature region is to include vegetable
Region.
Preferably, the area determination module includes：
Image splits submodule, for utilizing the example parted pattern based on depth convolutional neural networks pre-set,
Example segmentation is carried out to described first image, obtains label corresponding to each pixel, and the type of label；
Region division submodule, for the region that the type identical pixel of all labels forms to be defined as into first respectively
Characteristic area, and second feature region.
Preferably, described image adjusting module includes：
Recognition of face submodule, for carrying out recognition of face, and positioning key point to the fisrt feature region；
Character features detection sub-module, for utilizing the age detection depth convolutional neural networks model inspection pre-set
The age of personage in the fisrt feature region, and, utilize the sex-screening depth convolutional neural networks model pre-set
Detect the sex of personage in the fisrt feature region；
Personage's parameter acquiring submodule, for obtaining the special effect parameters that there is corresponding relation with the age, sex；
Personage's parameter adjustment submodule, for for the key point, the present image in the fisrt feature region to be joined
Number is changed to the special effect parameters.
Preferably, described image adjusting module also includes：
Vegetable identifies submodule, for being known using the vegetable identification model based on depth convolutional neural networks pre-set
The title of the vegetable in not described second feature region；
Vegetable parameter identifies submodule, for obtaining the filter params for having corresponding relation with the title of the vegetable；
Vegetable parameter adjustment submodule, for being imitated using the filter params to second feature region addition filter
Fruit.
Preferably, described image adjusting module also includes：
Personalized sentence acquisition submodule, for being obtained from preset personalized statement library with menu name with corresponding
The personalized sentence of relation；
Personalized sentence sets submodule, for adding the personalized sentence in default key point；
Image generates submodule, and described for the fisrt feature region after being operated based on adjustment, second feature region
Personalized sentence, generate the second image.
In order to solve the above problems, the embodiment of the invention discloses a kind of electronic equipment, including memory, processor and deposit
Storage realizes above-mentioned on a memory and the computer program that can run on a processor, during the computing device described program
The step of method of anticipating.
In order to solve the above problems, the embodiment of the invention discloses a kind of computer-readable recording medium, it is stored thereon with
Computer program, it is characterised in that the program realizes the step of above-mentioned any one method when being executed by processor.
The embodiment of the present invention includes advantages below：
In embodiments of the present invention, when receiving Client-initiated image acquisition commands, the first image, Ran Houji are gathered
Fisrt feature region and second feature region are determined from the first image in example segmentation, then fisrt feature region is entered respectively
The adjustment operation of row first, and, the second adjustment is carried out to second feature region and is operated, finally, will obtain and performed the first adjustment
The second picture showing after operation and the second adjustment operation is to user.Realize and the different target in same image is carried out respectively
Identification, and determine to include the regional of different target, then adjusted accordingly respectively for the different characteristics of regional
Whole operation.So, user need not be adjusted operation to image manually using third party application, improve user's body
Test, furthermore, it is possible to for carrying out differentiation adjustment respectively the characteristics of different zones in same image, rather than same image is entered
The unified adjustment of row, add the aesthetic feeling of image.
Brief description of the drawings
Fig. 1 is a kind of step flow chart of the embodiment of the method for image procossing of the present invention；
Fig. 2 is the distribution schematic diagram of each key point of the face of the present invention；
Fig. 3 is a kind of structured flowchart of the device embodiment of image procossing of the present invention.
Embodiment
In order to facilitate the understanding of the purposes, features and advantages of the present invention, it is below in conjunction with the accompanying drawings and specific real
Applying mode, the present invention is further detailed explanation.
Reference picture 1, shows a kind of step flow chart of the embodiment of the method for image procossing of the present invention, and methods described should
For mobile terminal.
In embodiments of the present invention, mobile terminal can have following features：
(1) on hardware systems, equipment possesses central processing unit, memory, input block and output block, that is,
Say, equipment is often the microcomputer devices for possessing communication function.Furthermore it is also possible to there are multiple input modes, such as key
Disk, mouse, touch-screen, transmitter and shooting are first-class, and can be adjusted input as needed.Meanwhile equipment often has
A variety of way of outputs, such as receiver, display screen, it can also be adjusted as needed；
(2) on software architecture, equipment must possess operating system, as Windows Mobile, Symbian, Palm,
Android, iOS etc..Meanwhile these operating systems are more and more open, of the operating system platform exploitation based on these openings
Property application program emerge in an endless stream, such as address book, schedule, notepad, calculator and various types of games are dramatically full
The foot demand of personalized user；
(3) in communication capacity, equipment has flexible access way and high-bandwidth communication performance, and can be selected by
The business selected and residing environment, the communication mode selected by adjust automatically, so as to be user-friendly.Equipment can support GSM,
WCDMA, CDMA2000, TDSCDMA, Wi-Fi and WiMAX etc., so as to adapt to multiple types network, not only support voice industry
Business, more supports a variety of wireless data services；
(4) in function using upper, equipment more focuses on hommization, personalization and multifunction.With computer technology
Development, equipment enter the pattern of " human-centred " from the pattern of " facility center management ", are integrated with embedding assembly, control skill
Art, artificial intelligence technology and biometrics etc., people-oriented objective is fully demonstrated.Due to the hair of software engineering
Exhibition, equipment can be adjusted according to demands of individuals and set, more personalized.Meanwhile equipment is integrated with numerous software and hardwares in itself,
Function is also stronger and stronger.
In a kind of preferred embodiment of the present invention, the application program specified is installed on the mobile terminal.Specifically, with
Family can use the application program that this is specified to buy the food and drink such as take-away, and the application journey specified by this in businessman by network
Ordered pair food and drink, businessman are evaluated, and anyone can see businessman, all evaluations of food and drink on network, wherein, evaluation can
Evaluation in a manner of including word evaluation, picture evaluation etc..
Described method specifically may include steps of：
Step 101, Client-initiated image acquisition commands are received, gather the first image；
, it is necessary to which image is taken on site when in the specific implementation, user is carrying out picture evaluation to food and drink, businessman, now, when
When mobile terminal receives user's initiation image acquisition commands, mobile terminal gathers the first image.
In a kind of preferred embodiment of the present invention, the reception Client-initiated image capture instruction, the first image is gathered
The step of include：
The image acquisition commands that user is initiated by the application program specified are received, gather the first image.
Specifically, user uses the application program specified, when carrying out picture evaluation to food and drink, businessman, can click on has
The virtual push button of shoot function, now, application program will call the camera of mobile terminal, and the current interface of mobile terminal is just
IMAQ interface can be jumped to from evaluation interface, user there can be the void of shoot function by clicking in IMAQ interface
Intend button and carry out IMAQ, IMAQ can also be carried out by the entity shooting push button of mobile terminal.
Certainly, user can also gather video except that can gather image, for example, IMAQ interface is adopted including image
Integrated mode, video capture mode both of which, when user selects image acquisition modality, what is collected is exactly static image,
When user selects video capture mode, what is collected is exactly dynamic video.
And virtual push button, or the mode of entity button certain time can be pinned to adopt by user by realizing video acquisition
Collection, virtual push button can also be clicked on for the first time by user, or entity button starts to gather video, user clicks on virtual for the second time
Button, or entity button terminate to gather the mode of video to gather, or also possible to gather video otherwise.
Step 102, fisrt feature region, and second feature region are determined from described first image；
" example segmentation " is a subdivision field in image segmentation, corresponding with " semantic segmentation ".So-called semantic segmentation,
As its name suggests, the semantic segmentation of pixel scale is to mark off corresponding classification to each pixel in image, that is, realizes pixel
The classification of rank.And the specific object of class, as example, then example splits the classification that not only carry out pixel scale, also needs
Different examples is differentiated on the basis of specific classification.That is, be to outline all objects in piece image to come, and will
Object carries out the segmentation extraction of pixel scale.
Such as image has first, second, the third three people, if carrying out semantic segmentation to image, then image can only be identified
Middle someone, when three people are to separate not overlapping, then three people can be identified respectively, when three people are overlapping, then only
Overall profile is can recognize that, cannot be distinguished by specifically there are several individuals, but if carrying out example segmentation to image, then no matter three
Whether people is overlapping, can identify specifically there are several individuals, and everyone profile.
The embodiment of the present invention is exactly to identify in the first image whether include portrait using above-mentioned principle, of the invention a kind of
In preferred embodiment, the fisrt feature region is the region comprising portrait.
When in the specific implementation, user is carrying out picture evaluation to food and drink, businessman, picture is possible to that portrait can be included.Example
Such as, companion is photographed picture and suffered, then just comprising " people " and " vegetable in picture by user together when shooting " pork braised in brown sauce "
(pork braised in brown sauce) ", then now just can just identify that image includes portrait using example segmentation, therefore, will include
The region of portrait is defined as fisrt feature region.
It should be noted that fisrt feature region does not imply that this includes the region of a people, and refer to contain people
Region, can calculate has three people in fisrt feature region, such as the first image, then fisrt feature region is just to include three
The region of individual portrait.
While whether fisrt feature region is included in judging the first image, whether it can judge in the first image comprising the
Two characteristic areas, in a kind of preferred embodiment of the present invention, the second feature region is the region comprising vegetable, wherein, dish
Product can be drink or food, for example, including coffee, and/or cake, and/or region of pork braised in brown sauce etc..
It should be noted that second feature region does not imply that this includes a kind of region of vegetable, and refer to contain
The region of one or more vegetables, it can calculate and coffee, cake, pork braised in brown sauce are included in second feature region, such as the first image,
So second feature region just be comprising coffee, cake, pork braised in brown sauce region.
It is described that fisrt feature region, Yi Ji are determined from described first image in a kind of preferred embodiment of the present invention
The step of two characteristic areas, includes：
Using the example parted pattern based on depth convolutional neural networks pre-set, described first image is carried out real
Example segmentation, obtains label corresponding to each pixel, and the type of label；
The region that the type identical pixel of all labels forms is defined as fisrt feature region, and the second spy respectively
Levy region.
Convolutional network is for identification two-dimensional shapes and a multilayer perceptron of particular design, and this network structure is to flat
Shifting, proportional zoom, inclination or the deformation of his common form have height consistency.These good performances are that network is having supervision
Learn under mode, the structure of network mainly has partially connected and weights to share two features, includes the constraint of following form：
1st, feature extraction.Each neuron obtains the defeated people of cynapse from the local acceptance region of last layer, thus forces it to carry
Take local feature.Once a feature is extracted, as long as it is approx remained relative to the position of other features,
Its exact position just becomes without so important.
2nd, Feature Mapping.Each computation layer of network is made up of multiple Feature Mappings, each Feature Mapping
It is plane form.Single neuron shares identical synaptic weight collection under the constraints in plane, and this structure type has
Following beneficial effect：A. translation invariance.B. the reduction (being realized by the way that weights are shared) of free parameter quantity.
3rd, sub-sample.Each convolutional layer is followed by a computation layer for realizing local average and sub-sample, thus feature
The resolution ratio of mapping reduces.This operation has the output for making Feature Mapping under the susceptibility of translation and the deformation of other forms
The effect of drop.
Convolutional neural networks are the neutral nets of a multilayer, and every layer is made up of multiple two dimensional surfaces, and each plane by
Multiple independent neuron compositions.
For example, the image of a 1000x1000 pixel, just there is 1,000,000 hidden neurons, then if their full connections
(each pixel of each hidden neuron connection figure picture), just there is 1000x1000x1000000=10^12 connection,
Namely 10^12 weighting parameter.But the space relationship of image is local, just as people is the receptive field by a part
Go to experience that external image is the same, and each neuron need not be done to global image experiences, each neuron only experiences local
Image-region, then in higher, these different local neurons of impression are integrated to the letter that can be obtained by the overall situation
Cease.So, our cans reduce the number of connection, that is, reduce the number for the weighting parameter that neutral net needs are trained
.If local receptor field is 10x10, each receptive field of hidden layer only needs to be connected with this 10x10 topography, so 1
Million hidden neurons just only have 100,000,000 connections, i.e. 10^8 parameter.Than reducing four 0 (orders of magnitude) originally, so
Training is got up just less laborious.
Using in embodiments of the present invention, to realize using depth convolutional neural networks Model Identification personage, vegetable, just need
Will first respectively to the age detection depth convolutional neural networks model pre-set, pre-set based on depth convolutional Neural net
The vegetable identification model of network, is trained.
The training data by Pixel-level mark is collected first, for each pixel progress classification mark of image, and
Different examples is made a distinction in mark, that is, the image of variety classes, the vegetable of different names is inputted into depth convolution
Neural network model.Then deep learning method, such as FCN (Fully Convolutional Networks, full volume are utilized
Product), (Fully Convolutional Instance-aware Semantic Segmentation, full convolution perceive FCIS
Semantic segmentation), Mask R-CNN (Mask Representation-Convolutional Neural Network, template generation
Table convolutional neural networks) etc., carry out the training of example parted pattern.When in use, the input of the first image is pre-set into depth
Convolutional neural networks model, the first image is entered using the example parted pattern based on depth convolutional neural networks pre-set
Row example is split, and obtains label corresponding to each pixel, and the type of label.In practical application, user in shooting
One image generally all without an only main body, so during matching, can match multiple main bodys, therefore, for not
With body region pixel, it is necessary to generate different labels, the region that label identical pixel is formed then is defined as the
One characteristic area, and second feature region.Wherein, the type of label can be divided into " people " and " inhuman ", such as, portrait area
The type of label of pixel can be " 1 ", rather than the type can of the label of the pixel of portrait area is " 0 ".
For example, the label of the pixel of portrait area can be " people ", the type of label is " 1 ", and the pixel in coffee region
Label can be " coffee ", the type of label is " 0 ", and the label of the pixel in cake region can be " cake ", the class of label
Type is " 0 ", for a certain image, depth convolutional neural networks model to the first images match and generate " people ", " coffee ",
" cake " three kinds of labels, then, the type of all labels is defined as fisrt feature area for the region that the pixel of " 1 " is formed
Domain, all labels are defined as second feature region for the region of " 0 ".
Step 103, the first adjustment is carried out to the fisrt feature region respectively to operate, and, to the second feature area
Domain carries out the second adjustment operation, obtains the second image after having performed the first adjustment operation and the second adjustment operation；
Behind the fisrt feature region and second feature region during the first image is determined, it is possible to special to first respectively
Levy region and second feature and region carries out landscaping treatment.
It should be noted that carrying out landscaping treatment to fisrt feature region, and landscaping treatment is carried out to second feature region
It is independent, the first adjustment operation is carried out to fisrt feature region will not be applied to second feature region, to second feature region
Fisrt feature region will not be also applied to by carrying out the second adjustment operation.
In a kind of preferred embodiment of the present invention, the depth convolutional neural networks model includes age detection depth convolution
Neural network model, and sex-screening depth convolutional neural networks model；It is described that the first tune is carried out to the fisrt feature region
The step of whole operation, includes：
The fisrt feature regions are carried out with recognition of face, and positioning key point；
Utilize personage in fisrt feature region described in the age detection depth convolutional neural networks model inspection pre-set
Age, and, utilize personage in fisrt feature region described in the sex-screening depth convolutional neural networks model inspection
Sex；
Obtain the special effect parameters that there is corresponding relation with the age, sex；
It is the special effect parameters by the present image parameter change in the fisrt feature region for the key point.
In embodiments of the present invention, depth convolutional neural networks model includes age detection depth convolutional neural networks mould
Type, and sex-screening depth convolutional neural networks model, the two is two and separates independent model.Rolled up to age detection depth
When product neural network model is trained, the labels such as age, the sex of personage can be set in every character image of input,
Such as image X age, sex parameter are (0,30), wherein, " 0 " represents male, and " 30 " represented the age as 30 years old, image Y's
Age, sex parameter are (1,18), wherein, " 1 " represents schoolgirl, and " 18 " represented the age as 18 years old.Then depth convolution net is utilized
Road model, such as VGG (visual geometry group, two-dimensional bodies identification), AlexNet scheduling algorithms are examined to the age respectively
Depth measurement degree convolutional neural networks model, and sex-screening depth convolutional neural networks model are trained.For example, when in use,
First image is input in the sex-screening depth convolutional neural networks model trained, sex-screening depth convolution god
A two-dimentional vector can be obtained through network model, represents probability size of the image as masculinity and femininity respectively, chooses probability
Maximum classification is the tag along sort of the image, you can judges the image for male or women, the detection method at age is similarly.
After age of personage, sex is detected, it is possible to obtained from preset special effect parameters database with the age,
Sex has the special effect parameters of corresponding relation.Wherein, different sexes can be stored in special effect parameters database, the age corresponds
Special effect parameters, such as special effect parameters corresponding to 18 years old women are exactly different from special effect parameters corresponding to 40 years old women.Then
The parameter current of people's object area is changed to corresponding special effect parameters, so as to realize the beautification to personage.
And then can be with each key point of locating human face, as shown in Figure 2 to the progress recognition of face of people's object area.
It is described that the step of the second adjustment operates is carried out to the second feature region in a kind of preferred embodiment of the present invention
Including：
Utilize the vegetable in second feature region described in the vegetable identification depth convolutional neural networks Model Identification pre-set
Title；
Obtain the filter params that there is corresponding relation with the title of the vegetable；
Filtering effects are added to the second feature region using the filter params；
Specifically, when training vegetable disaggregated model, it usually needs the image of the vegetable of a large amount of classifications is as training image.
The form of the training image of model based on depth convolutional neural networks is usually (label, image), wherein, label is figure
The class label of picture, image are vegetable image, the title corresponding to every kind of vegetable image, wherein, vegetable can include drink
Product, food etc. are classified.Depth convolutional neural networks are identified by learning to different classes of image based on vegetable, training
The parameter of disaggregated model.When in use, other images to be recognized can be classified using the disaggregated model that training obtains,
The classification corresponding to the image is obtained, and the title being somebody's turn to do.Then the name with vegetable is obtained from preset filter database
Weighing-appliance has the filter params of corresponding relation, finally adds filtering effects to vegetable region using the filter params obtained.Wherein, filter
The one-to-one filtering effects of different classes of vegetable are stored with mirror database, and/or the vegetable of different names corresponds
Filtering effects, for example, noodles classification selects the filtering effects of delicious style, drink classification selects the filter effect of refrigerant style
Fruit, sweets classification select filtering effects of fragrant and sweet style etc..
In a kind of preferred embodiment of the present invention, the acquisition has performed the first adjustment operation and the second adjustment and operated
The step of rear the second image, includes：
The personalized sentence that there is corresponding relation with menu name is obtained from preset personalized statement library；
The personalized sentence is added in default key point；
Fisrt feature region, second feature region after being operated based on adjustment, and the personalized sentence, generation second
Image；
In the specific implementation, to be stored with different classes of vegetable in preset personalized statement library personalized correspondingly
Sentence, and/or the vegetable personalized sentence correspondingly of different names, can be with after the title or classification of vegetable is recognized
The personalized sentence that there is corresponding relation with menu name is obtained from personalized statement library, then near default key point
The personalized sentence is added, wherein, default key point is any point in Fig. 2.Finally, based on first after special effect processing
Second feature region after characteristic area, addition filtering effects, and personalized sentence, generate the second image.
Certainly, can also be by the first image in addition to any point in Fig. 2 is arranged into default key point
Other positions, such as some point in second feature region is set to default key point, also possible, the embodiment of the present invention pair
This is not restricted.
Step 104, second image is showed.
In a kind of preferred embodiment of the present invention, described the step of showing second image, includes：
Second image is showed by the application program specified.
Specifically, after the application program specified carries out a series of the second image of processing generations to the first image, by generation
Second picture showing is to user.That is, user is by virtual shooting push button or entity shooting push button shooting image, it is actual
What is seen is not the first image, but a series of the second figure that the application program specified generates to the first image after processing
Picture.
In embodiments of the present invention, when receiving Client-initiated image acquisition commands, the first image, Ran Houji are gathered
Fisrt feature region and second feature region are determined from the first image in example segmentation, then fisrt feature region is entered respectively
The adjustment operation of row first, and, the second adjustment is carried out to second feature region and is operated, finally, will obtain and performed the first adjustment
The second picture showing after operation and the second adjustment operation is to user.Realize and the different target in same image is carried out respectively
Identification, and determine to include the regional of different target, then adjusted accordingly respectively for the different characteristics of regional
Whole operation.So, user need not be adjusted operation to image manually using third party application, improve user's body
Test, furthermore, it is possible to for carrying out differentiation adjustment respectively the characteristics of different zones in same image, rather than same image is entered
The unified adjustment of row, add the aesthetic feeling of image.
It should be noted that for embodiment of the method, in order to be briefly described, therefore it is all expressed as to a series of action group
Close, but those skilled in the art should know, the embodiment of the present invention is not limited by described sequence of movement, because according to
According to the embodiment of the present invention, some steps can use other orders or carry out simultaneously.Secondly, those skilled in the art also should
Know, embodiment described in this description belongs to preferred embodiment, and the involved action not necessarily present invention is implemented
Necessary to example.
Reference picture 3, a kind of structured flowchart of the device embodiment of image procossing of the present invention is shown, can specifically be included
Following module：
Acquisition module 301, for receiving Client-initiated image acquisition commands, gather the first image；
Area determination module 302, for determining fisrt feature region, and second feature area from described first image
Domain；
Image adjustment module 303, operated for carrying out the first adjustment to the fisrt feature region respectively, and, to institute
State second feature region and carry out the second adjustment operation, obtain the performed after the first adjustment operation and the second adjustment operate
Two images；
Display module 304, for showing second image.
In a kind of preferred embodiment of the present invention, including the application journey specified is installed on the mobile terminal of described device
Sequence；
The acquisition module is additionally operable to receive the image acquisition commands that user is initiated by the application program specified, and adopts
Collect the first image；
The display module is additionally operable to show second image by the application program specified.
In a kind of preferred embodiment of the present invention, the fisrt feature region is the region comprising portrait, and described second is special
Sign region is the region comprising vegetable.
In a kind of preferred embodiment of the present invention, the area determination module includes：
Image splits submodule, for utilizing the example parted pattern based on depth convolutional neural networks pre-set,
Example segmentation is carried out to described first image, obtains label corresponding to each pixel, and the type of label；
Region division submodule, for the region that the type identical pixel of all labels forms to be defined as into first respectively
Characteristic area, and second feature region.
In a kind of preferred embodiment of the present invention, described image adjusting module includes：
Recognition of face submodule, for carrying out recognition of face, and positioning key point to the fisrt feature region；
Character features detection sub-module, for utilizing the age detection depth convolutional neural networks model inspection pre-set
The age of personage in the fisrt feature region, and, utilize the sex-screening depth convolutional neural networks model pre-set
Detect the sex of personage in the fisrt feature region；
Personage's parameter acquiring submodule, for obtaining the special effect parameters that there is corresponding relation with the age, sex；
Personage's parameter adjustment submodule, for for the key point, the present image in the fisrt feature region to be joined
Number is changed to the special effect parameters.
In a kind of preferred embodiment of the present invention, described image adjusting module also includes：
Vegetable identifies submodule, for being known using the vegetable identification model based on depth convolutional neural networks pre-set
The title of the vegetable in not described second feature region；
Vegetable parameter identifies submodule, for obtaining the filter params for having corresponding relation with the title of the vegetable；
Vegetable parameter adjustment submodule, for being imitated using the filter params to second feature region addition filter
Fruit.
In a kind of preferred embodiment of the present invention, described image adjusting module also includes：
Personalized sentence acquisition submodule, for being obtained from preset personalized statement library with menu name with corresponding
The personalized sentence of relation；
Personalized sentence sets submodule, for adding the personalized sentence in default key point；
Image generates submodule, and described for the fisrt feature region after being operated based on adjustment, second feature region
Personalized sentence, generate the second image.
The embodiments of the invention provide a kind of electronic equipment and a kind of computer-readable recording medium, the electronic equipment bag
The computer program that includes memory, processor and storage on a memory and can run on a processor, the computing device
Following steps are realized during described program：
Client-initiated image acquisition commands are received, gather the first image；
Fisrt feature region, and second feature region are determined from described first image；
The first adjustment is carried out to the fisrt feature region respectively to operate, and, the is carried out to the second feature region
Two adjustment operations, obtain the second image after having performed the first adjustment operation and the second adjustment operation；
Show second image.
Preferably, the application program specified is installed on the mobile terminal；
It is described reception Client-initiated image capture instruction, gather the first image the step of include：
The image acquisition commands that user is initiated by the application program specified are received, gather the first image；
Described the step of showing second image, includes：
Second image is showed by the application program specified.
Preferably, the fisrt feature region is the region comprising portrait, and the second feature region is to include vegetable
Region.
Preferably, it is described to determine that fisrt feature region, and the step of second feature region are wrapped from described first image
Include：
Using the example parted pattern based on depth convolutional neural networks pre-set, described first image is carried out real
Example segmentation, obtains label corresponding to each pixel, and the type of label；
The region that the type identical pixel of all labels forms is defined as fisrt feature region, and the second spy respectively
Levy region.
Preferably, described the step of the first adjustment operation is carried out to the fisrt feature region, includes：
Recognition of face, and positioning key point are carried out to the fisrt feature region；
Utilize personage in fisrt feature region described in the age detection depth convolutional neural networks model inspection pre-set
Age, and, using in fisrt feature region described in the sex-screening depth convolutional neural networks model inspection pre-set
The sex of personage；
Obtain the special effect parameters that there is corresponding relation with the age, sex；
It is the special effect parameters by the present image parameter change in the fisrt feature region for the key point.
Preferably, described the step of the second adjustment operation is carried out to the second feature region, includes：
The second feature region is identified using the vegetable identification model based on depth convolutional neural networks pre-set
Vegetable title；
Obtain the filter params that there is corresponding relation with the title of the vegetable；
Filtering effects are added to the second feature region using the filter params.
Preferably, the step of obtaining the second image after having performed the first adjustment operation and the second adjustment operation is wrapped
Include：
The personalized sentence that there is corresponding relation with menu name is obtained from preset personalized statement library；
The personalized sentence is added in default key point；
Fisrt feature region, second feature region after being operated based on adjustment, and the personalized sentence, generation second
Image.
The computer-readable recording medium storage has computer program, can be realized such as when the program is executed by processor
Lower step：
Client-initiated image acquisition commands are received, gather the first image；
Fisrt feature region, and second feature region are determined from described first image；
The first adjustment is carried out to the fisrt feature region respectively to operate, and, the is carried out to the second feature region
Two adjustment operations, obtain the second image after having performed the first adjustment operation and the second adjustment operation；
Show second image.
Preferably, the application program specified is installed on the mobile terminal；
It is described reception Client-initiated image capture instruction, gather the first image the step of include：
The image acquisition commands that user is initiated by the application program specified are received, gather the first image；
Described the step of showing second image, includes：
Second image is showed by the application program specified.
Preferably, the fisrt feature region is the region comprising portrait, and the second feature region is to include vegetable
Region.
Preferably, it is described to determine that fisrt feature region, and the step of second feature region are wrapped from described first image
Include：
Using the example parted pattern based on depth convolutional neural networks pre-set, described first image is carried out real
Example segmentation, obtains label corresponding to each pixel, and the type of label；
The region that the type identical pixel of all labels forms is defined as fisrt feature region, and the second spy respectively
Levy region.
Preferably, described the step of the first adjustment operation is carried out to the fisrt feature region, includes：
Recognition of face, and positioning key point are carried out to the fisrt feature region；
Utilize personage in fisrt feature region described in the age detection depth convolutional neural networks model inspection pre-set
Age, and, using in fisrt feature region described in the sex-screening depth convolutional neural networks model inspection pre-set
The sex of personage；
Obtain the special effect parameters that there is corresponding relation with the age, sex；
It is the special effect parameters by the present image parameter change in the fisrt feature region for the key point.
Preferably, described the step of the second adjustment operation is carried out to the second feature region, includes：
The second feature region is identified using the vegetable identification model based on depth convolutional neural networks pre-set
Vegetable title；
Obtain the filter params that there is corresponding relation with the title of the vegetable；
Filtering effects are added to the second feature region using the filter params.
Preferably, the step of obtaining the second image after having performed the first adjustment operation and the second adjustment operation is wrapped
Include：
The personalized sentence that there is corresponding relation with menu name is obtained from preset personalized statement library；
The personalized sentence is added in default key point；
Fisrt feature region, second feature region after being operated based on adjustment, and the personalized sentence, generation second
Image.
For device embodiment, because it is substantially similar to embodiment of the method, so description is fairly simple, it is related
Part illustrates referring to the part of embodiment of the method.
Each embodiment in this specification is described by the way of progressive, what each embodiment stressed be with
The difference of other embodiment, between each embodiment identical similar part mutually referring to.
It should be understood by those skilled in the art that, the embodiment of the embodiment of the present invention can be provided as method, apparatus or calculate
Machine program product.Therefore, the embodiment of the present invention can use complete hardware embodiment, complete software embodiment or combine software and
The form of the embodiment of hardware aspect.Moreover, the embodiment of the present invention can use one or more wherein include computer can
With in the computer-usable storage medium (including but is not limited to magnetic disk storage, CD-ROM, optical memory etc.) of program code
The form of the computer program product of implementation.
The embodiment of the present invention is with reference to method according to embodiments of the present invention, terminal device (system) and computer program
The flow chart and/or block diagram of product describes.It should be understood that can be by computer program instructions implementation process figure and/or block diagram
In each flow and/or square frame and the flow in flow chart and/or block diagram and/or the combination of square frame.These can be provided
Computer program instructions are set to all-purpose computer, special-purpose computer, Embedded Processor or other programmable data processing terminals
Standby processor is to produce a machine so that is held by the processor of computer or other programmable data processing terminal equipments
Capable instruction is produced for realizing in one flow of flow chart or multiple flows and/or one square frame of block diagram or multiple square frames
The device for the function of specifying.
These computer program instructions, which may be alternatively stored in, can guide computer or other programmable data processing terminal equipments
In the computer-readable memory to work in a specific way so that the instruction being stored in the computer-readable memory produces bag
The manufacture of command device is included, the command device is realized in one flow of flow chart or multiple flows and/or one side of block diagram
The function of being specified in frame or multiple square frames.
These computer program instructions can be also loaded into computer or other programmable data processing terminal equipments so that
Series of operation steps is performed on computer or other programmable terminal equipments to produce computer implemented processing, so that
The instruction performed on computer or other programmable terminal equipments is provided for realizing in one flow of flow chart or multiple flows
And/or specified in one square frame of block diagram or multiple square frames function the step of.
Although having been described for the preferred embodiment of the embodiment of the present invention, those skilled in the art once know base
This creative concept, then other change and modification can be made to these embodiments.So appended claims are intended to be construed to
Including preferred embodiment and fall into having altered and changing for range of embodiment of the invention.
Finally, it is to be noted that, herein, such as first and second or the like relational terms be used merely to by
One entity or operation make a distinction with another entity or operation, and not necessarily require or imply these entities or operation
Between any this actual relation or order be present.Moreover, term " comprising ", "comprising" or its any other variant meaning
Covering including for nonexcludability, so that process, method, article or terminal device including a series of elements are not only wrapped
Those key elements, but also the other element including being not expressly set out are included, or is also included for this process, method, article
Or the key element that terminal device is intrinsic.In the absence of more restrictions, wanted by what sentence "including a ..." limited
Element, it is not excluded that other identical element in the process including the key element, method, article or terminal device also be present.
Method to a kind of image procossing provided by the present invention and a kind of device of image procossing above, carry out in detail
Introduce, specific case used herein is set forth to the principle and embodiment of the present invention, the explanation of above example
It is only intended to help the method and its core concept for understanding the present invention；Meanwhile for those of ordinary skill in the art, according to this
The thought of invention, there will be changes in specific embodiments and applications, in summary, this specification content should
It is interpreted as limitation of the present invention.
Claims (10)
- A kind of 1. method of image procossing, it is characterised in that methods described is applied to mobile terminal, and described method includes：Client-initiated image acquisition commands are received, gather the first image；Fisrt feature region, and second feature region are determined from described first image；The first adjustment is carried out to the fisrt feature region respectively to operate, and, second is carried out to the second feature region and is adjusted Whole operation, obtain the second image after having performed the first adjustment operation and the second adjustment operation；Show second image.
- 2. according to the method for claim 1, it is characterised in that the application program specified is installed on the mobile terminal；It is described reception Client-initiated image capture instruction, gather the first image the step of include：The image acquisition commands that user is initiated by the application program specified are received, gather the first image；Described the step of showing second image, includes：Second image is showed by the application program specified.
- 3. according to the method for claim 1, it is characterised in that the fisrt feature region is the region comprising portrait, institute It is the region comprising vegetable to state second feature region.
- 4. according to the method for claim 1, it is characterised in that described that fisrt feature area is determined from described first image Domain, and the step of second feature region include：Using the example parted pattern based on depth convolutional neural networks pre-set, example point is carried out to described first image Cut, obtain label corresponding to each pixel, and the type of label；The region that the type identical pixel of all labels forms is defined as fisrt feature region, and second feature area respectively Domain.
- 5. according to the method for claim 1, it is characterised in that described that the first adjustment behaviour is carried out to the fisrt feature region As the step of include：Recognition of face, and positioning key point are carried out to the fisrt feature region；Utilize the year of personage in fisrt feature region described in the age detection depth convolutional neural networks model inspection pre-set Age, and, utilize personage in fisrt feature region described in the sex-screening depth convolutional neural networks model inspection pre-set Sex；Obtain the special effect parameters that there is corresponding relation with the age, sex；It is the special effect parameters by the present image parameter change in the fisrt feature region for the key point.
- 6. according to the method for claim 1, it is characterised in that described that the second adjustment behaviour is carried out to the second feature region As the step of include：The dish in the second feature region is identified using the vegetable identification model based on depth convolutional neural networks pre-set The title of product；Obtain the filter params that there is corresponding relation with the title of the vegetable；Filtering effects are added to the second feature region using the filter params.
- 7. according to the method described in claim any one of 1-6, it is characterised in that acquisition has performed it is described first adjustment operation with Second includes the step of adjusting the second image after operation：The personalized sentence that there is corresponding relation with menu name is obtained from preset personalized statement library；The personalized sentence is added in default key point；Fisrt feature region, second feature region after being operated based on adjustment, and the personalized sentence, generate the second figure Picture.
- A kind of 8. device of image procossing, it is characterised in that including：Acquisition module, for receiving Client-initiated image acquisition commands, gather the first image；Area determination module, for determining fisrt feature region, and second feature region from described first image；Image adjustment module, operated for carrying out the first adjustment to the fisrt feature region respectively, and, it is special to described second Levy region and carry out the second adjustment operation, obtain the second image after having performed the first adjustment operation and the second adjustment operation；Display module, for showing second image.
- 9. a kind of electronic equipment, the equipment includes memory and processor, and being stored with the memory can be at the place The computer program run on reason device, it is characterised in that the processor performs aforesaid right when running the computer program It is required that the method described in 1 to 7 any one.
- 10. a kind of computer-readable recording medium, is stored thereon with computer program, it is characterised in that the computer program The method described in any one of the claims 1 to 7 is performed when being run by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710853951.7A CN107798653B (en) | 2017-09-20 | 2017-09-20 | Image processing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710853951.7A CN107798653B (en) | 2017-09-20 | 2017-09-20 | Image processing method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107798653A true CN107798653A (en) | 2018-03-13 |
| CN107798653B CN107798653B (en) | 2019-12-24 |
Family
ID=61531963
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710853951.7A Active CN107798653B (en) | 2017-09-20 | 2017-09-20 | Image processing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107798653B (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108629319A (en) * | 2018-05-09 | 2018-10-09 | 北京嘀嘀无限科技发展有限公司 | Image detecting method and system |
| CN108764370A (en) * | 2018-06-08 | 2018-11-06 | Oppo广东移动通信有限公司 | Image processing method, device, computer readable storage medium and computer equipment |
| CN108764051A (en) * | 2018-04-28 | 2018-11-06 | Oppo广东移动通信有限公司 | Image processing method, device and mobile terminal |
| CN108776819A (en) * | 2018-06-05 | 2018-11-09 | Oppo广东移动通信有限公司 | A kind of target identification method, mobile terminal and computer readable storage medium |
| CN108805095A (en) * | 2018-06-19 | 2018-11-13 | Oppo广东移动通信有限公司 | image processing method, device, mobile terminal and computer readable storage medium |
| CN108846314A (en) * | 2018-05-08 | 2018-11-20 | 天津大学 | A kind of food materials identification system and food materials discrimination method based on deep learning |
| CN108875676A (en) * | 2018-06-28 | 2018-11-23 | 北京旷视科技有限公司 | Biopsy method, apparatus and system |
| CN108898587A (en) * | 2018-06-19 | 2018-11-27 | Oppo广东移动通信有限公司 | Image processing method, picture processing unit and terminal device |
| CN109035147A (en) * | 2018-08-10 | 2018-12-18 | Oppo广东移动通信有限公司 | Image processing method and device, electronic device, storage medium and computer equipment |
| CN109117701A (en) * | 2018-06-05 | 2019-01-01 | 东南大学 | Pedestrian's intension recognizing method based on picture scroll product |
| CN109814551A (en) * | 2019-01-04 | 2019-05-28 | 丰疆智慧农业股份有限公司 | Cereal handles automated driving system, automatic Pilot method and automatic identifying method |
| CN110298405A (en) * | 2019-07-03 | 2019-10-01 | 北京字节跳动网络技术有限公司 | Classification recognition methods and device, storage medium and terminal |
| CN110826504A (en) * | 2019-11-07 | 2020-02-21 | 深圳市微埃智能科技有限公司 | Visual identification method and device combined with fine-grained classification and computer equipment |
| CN111340912A (en) * | 2020-02-20 | 2020-06-26 | 北京市商汤科技开发有限公司 | Scene graph generation method and device and storage medium |
| CN112840376A (en) * | 2018-10-15 | 2021-05-25 | 华为技术有限公司 | Image processing method, device and equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103578098A (en) * | 2012-08-07 | 2014-02-12 | 阿里巴巴集团控股有限公司 | Method and device for extracting commodity body in commodity picture |
| CN103761241A (en) * | 2013-12-13 | 2014-04-30 | 北京奇虎科技有限公司 | Method for uploading picture data and browser |
| CN104794462A (en) * | 2015-05-11 | 2015-07-22 | 北京锤子数码科技有限公司 | Figure image processing method and device |
| CN106959759A (en) * | 2017-03-31 | 2017-07-18 | 联想(北京)有限公司 | A kind of data processing method and device |
-
2017
- 2017-09-20 CN CN201710853951.7A patent/CN107798653B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103578098A (en) * | 2012-08-07 | 2014-02-12 | 阿里巴巴集团控股有限公司 | Method and device for extracting commodity body in commodity picture |
| CN103761241A (en) * | 2013-12-13 | 2014-04-30 | 北京奇虎科技有限公司 | Method for uploading picture data and browser |
| CN104794462A (en) * | 2015-05-11 | 2015-07-22 | 北京锤子数码科技有限公司 | Figure image processing method and device |
| CN106959759A (en) * | 2017-03-31 | 2017-07-18 | 联想(北京)有限公司 | A kind of data processing method and device |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108764051A (en) * | 2018-04-28 | 2018-11-06 | Oppo广东移动通信有限公司 | Image processing method, device and mobile terminal |
| CN108764051B (en) * | 2018-04-28 | 2021-07-13 | Oppo广东移动通信有限公司 | Image processing method and device and mobile terminal |
| CN108846314A (en) * | 2018-05-08 | 2018-11-20 | 天津大学 | A kind of food materials identification system and food materials discrimination method based on deep learning |
| CN108629319A (en) * | 2018-05-09 | 2018-10-09 | 北京嘀嘀无限科技发展有限公司 | Image detecting method and system |
| CN108629319B (en) * | 2018-05-09 | 2020-01-07 | 北京嘀嘀无限科技发展有限公司 | Image detection method and system |
| CN109117701A (en) * | 2018-06-05 | 2019-01-01 | 东南大学 | Pedestrian's intension recognizing method based on picture scroll product |
| CN109117701B (en) * | 2018-06-05 | 2022-01-28 | 东南大学 | Pedestrian intention identification method based on graph convolution |
| CN108776819A (en) * | 2018-06-05 | 2018-11-09 | Oppo广东移动通信有限公司 | A kind of target identification method, mobile terminal and computer readable storage medium |
| CN108764370B (en) * | 2018-06-08 | 2021-03-12 | Oppo广东移动通信有限公司 | Image processing method, image processing device, computer-readable storage medium and computer equipment |
| CN108764370A (en) * | 2018-06-08 | 2018-11-06 | Oppo广东移动通信有限公司 | Image processing method, device, computer readable storage medium and computer equipment |
| CN108898587A (en) * | 2018-06-19 | 2018-11-27 | Oppo广东移动通信有限公司 | Image processing method, picture processing unit and terminal device |
| CN108805095A (en) * | 2018-06-19 | 2018-11-13 | Oppo广东移动通信有限公司 | image processing method, device, mobile terminal and computer readable storage medium |
| CN108875676A (en) * | 2018-06-28 | 2018-11-23 | 北京旷视科技有限公司 | Biopsy method, apparatus and system |
| US11195037B2 (en) | 2018-06-28 | 2021-12-07 | Beijing Kuangshi Technology Co., Ltd. | Living body detection method and system, computer-readable storage medium |
| CN109035147A (en) * | 2018-08-10 | 2018-12-18 | Oppo广东移动通信有限公司 | Image processing method and device, electronic device, storage medium and computer equipment |
| CN109035147B (en) * | 2018-08-10 | 2023-12-12 | Oppo广东移动通信有限公司 | Image processing method and device, electronic device, storage medium and computer equipment |
| CN112840376A (en) * | 2018-10-15 | 2021-05-25 | 华为技术有限公司 | Image processing method, device and equipment |
| CN109814551A (en) * | 2019-01-04 | 2019-05-28 | 丰疆智慧农业股份有限公司 | Cereal handles automated driving system, automatic Pilot method and automatic identifying method |
| CN110298405A (en) * | 2019-07-03 | 2019-10-01 | 北京字节跳动网络技术有限公司 | Classification recognition methods and device, storage medium and terminal |
| CN110826504A (en) * | 2019-11-07 | 2020-02-21 | 深圳市微埃智能科技有限公司 | Visual identification method and device combined with fine-grained classification and computer equipment |
| CN111340912A (en) * | 2020-02-20 | 2020-06-26 | 北京市商汤科技开发有限公司 | Scene graph generation method and device and storage medium |
| CN111340912B (en) * | 2020-02-20 | 2022-12-23 | 北京市商汤科技开发有限公司 | Scene graph generation method and device and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107798653B (en) | 2019-12-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107798653A (en) | A kind of method of image procossing and a kind of device | |
| CN111898709B (en) | Image classification method and device | |
| KR102102161B1 (en) | Method, apparatus and computer program for extracting representative feature of object in image | |
| CN104715023B (en) | Method of Commodity Recommendation based on video content and system | |
| US9734426B2 (en) | Automated food recognition and nutritional estimation with a personal mobile electronic device | |
| CN110222787A (en) | Multiscale target detection method, device, computer equipment and storage medium | |
| US8577962B2 (en) | Server apparatus, client apparatus, content recommendation method, and program | |
| CN111104898A (en) | Image scene classification method and device based on target semantics and attention mechanism | |
| CN107194898A (en) | The method for pushing of the methods of exhibiting of hotel's image, storage medium and hotel information | |
| CN108921058A (en) | Fish identification method, medium, terminal device and device based on deep learning | |
| CN111696137B (en) | Target tracking method based on multilayer feature mixing and attention mechanism | |
| CN107145845A (en) | The pedestrian detection method merged based on deep learning and multi-characteristic points | |
| CN111491187B (en) | Video recommendation method, device, equipment and storage medium | |
| CN111079623A (en) | Target detection method, device and storage medium | |
| CN110110719A (en) | A kind of object detection method based on attention layer region convolutional neural networks | |
| CN104951770B (en) | Construction method, application process and the related device of face database | |
| CN110263768A (en) | A kind of face identification method based on depth residual error network | |
| CN108664946A (en) | Stream of people's characteristic-acquisition method based on image and device | |
| CN110097090A (en) | A kind of image fine granularity recognition methods based on multi-scale feature fusion | |
| CN107463906A (en) | The method and device of Face datection | |
| CN107145839A (en) | A kind of fingerprint image completion analogy method and its system | |
| CN116863539A (en) | Fall figure target detection method based on optimized YOLOv8s network structure | |
| CN109636764A (en) | A kind of image style transfer method based on deep learning and conspicuousness detection | |
| CN110599455A (en) | Display screen defect detection network model, method and device, electronic equipment and storage medium | |
| CN106649629A (en) | System connecting books with electronic resources |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |